국가수준 학업성취도 평가의 표집설계 개선 방안 연구

Abstract

The purpose of the present study is to improve the sampling design of the National Assessment of Educational Achievement (NAEA) to enhance the representativeness of the sampled data and the accuracy of statistical estimates. Based on the census data of 2016 NAEA, we conducted a simulation study by varying sampling methods, sample size, explicit stratification variable, and the expected class size, and compared the results from sampled data with the census results. The study found the appropriateness of stratified sampling relative to cluster sampling and efficiency of 3% sampling, and suggests revision of the expected class size reflecting the real educational settings, and inclusion of area size as an explicit stratification to improve the representation of the sampled data. The study implications can be used for improving sampling design in large scale assessments.